Pith. sign in

REVIEW 3 major objections 4 minor 65 references

Addressing Missing Data Issue for Diffusion-based Recommendation

T0 review · 3 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Deliberately deleting extra items from a user's history makes diffusion recommenders tolerate the gaps that are already there, and a Thompson-sampling edit rule keeps the preference trajectory intact.

desk verdict Novel and well-tested idea for diffusion recommenders, but the theoretical 'proof' is a circular assumption; the empirical results carry the paper. read the letter →

arxiv 2505.12283 v1 pith:S53RUHGX submitted 2025-05-18 cs.IR

Wenyu Mao , Zhengyi Yang , Jiancan Wu , Haozhe Liu , Yancheng Yuan , Xiang Wang , Xiangnan He This is my paper · ORCID
classification cs.IR
keywords diffusionmodelssequentialrecommendationmissingdataThompsonsamplingconsistencyregularizationdenoisingimplicitgenerative
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Diffusion-based sequential recommenders generate the next item by denoising under guidance from a user's observed history, so missing entries in that history weaken the generated item. This paper argues that recovering the missing entries is impractical because missingness is uncertain in both occurrence and content, and instead proposes TDM, which simulates additional missing data by deliberately removing items from the guidance sequence. The removals are chosen by dual-side Thompson sampling using local continuity and global stability scores, so the user-preference trajectory stays approximately unchanged while the sequence acquires gaps like the ones the model should tolerate. Training the diffusion model to reconstruct the next item from both observed and edited sequences acts as consistency regularization, and a Taylor-expansion argument says that robustness to simulated gaps transfers to real gaps by extrapolation. The paper reports consistent gains over prior sequential and generative recommenders on six datasets.

What carries the argument

The load-bearing mechanism is dual-side Thompson sampling built from two probability models. Locally, each item $n$ in a sequence gets a continuity score $\mathrm{con}_n=\exp(\mathrm{sim}(\mathbf e_n,\mathbf e_{n+1}))/\sum_{n'}\exp(\mathrm{sim}(\mathbf e_{n'},\mathbf e_{n'+1}))$; globally, each sequence $k$ gets a stability score $\mathrm{sta}_k$ obtained by softmax-normalizing the entropy of its continuity scores. These scores parameterize two Thompson-sampling distributions, and thresholds $\lambda_1,\lambda_2$ decide whether a sequence is edited and whether an item is replaced by a dummy token. The edited sequence is encoded into guidance $\mathbf g$, the denoising network predicts the next-item embedding directly, and the reconstruction loss doubles as consistency regularization because observed and edited versions are perturbed pairs. The extrapolation proof applies a first-order Taylor expansion to show $\|f_\theta(\cdot,\bar{\mathbf g},\cdot)-f_\theta(\cdot,\hat{\mathbf g},\cdot)\| \le C\|f_\theta(\cdot,\hat{\mathbf g},\cdot)-f_\theta(\cdot,\tilde{\mathbf g},\cdot)\|$ whenever the simulated missingness mechanism aligns with the real one.

What would settle it

A controlled experiment can test the extrapolation claim directly: generate test-set missingness by a mechanism opposite to DTS, such as deleting low-continuity items or hiding items that were never exposed, and compare TDM with a no-simulation diffusion recommender such as DreamRec. If TDM's advantage shrinks or reverses under such missingness, the claim that consistency on simulated gaps transfers to real gaps is falsified; a positive result would require TDM to keep its gain across random, preference-preserving, and exposure-based missingness.

Watch

Extended reading notes

Core claim

The paper's central claim is that the right response to missing interaction history is not to repair it but to make the recommender indifferent to gaps, by injecting simulated missingness during training. TDM's dual-side Thompson sampling computes a continuity score for every adjacent pair (softmax-normalized cosine similarity) and an entropy-based stability score for each whole sequence, then drops high-continuity items from high-stability sequences with probabilities governed by those scores; replacing an item with a dummy token and re-encoding the edited sequence by a Transformer yields the guidance for the diffusion denoiser. Because the reconstruction loss also bounds the difference between predictions made from observed and edited guidance, the model becomes invariant to preference-preserving perturbations, and the paper's extrapolation argument extends that invariance to real missing data. Empirically, TDM improves over the best baselines in hit ratio and NDCG on YooChoose, KuaiRec, Zhihu, Steam, Beauty, and Toys, and applying DTS to non-diffusion recommenders yields gains in most settings.

Load-bearing premise

The whole robustness argument rests on the assumption that real missing data in user histories behaves like the preference-preserving gaps DTS creates, so that the guidance differences from real and simulated missingness have comparable direction and size; if real gaps come from exposure bias, popularity, or other non-preference mechanisms, the extrapolation guarantee does not follow.

Editorial extensions

If this is right

  • DTS is a model-agnostic editing layer: it can be attached to any recommender that encodes a history sequence, and the paper shows gains on GRU4Rec, SASRec, CL4SRec, Caser, AdaRanker, and DiffuASR as well as on diffusion recommenders.
  • Recovering-based pipelines such as imputation or insertion are not necessary for missing-data robustness; simulating gaps plus consistency regularization is cheaper and avoids the risk of injecting guessed items.
  • The consistency-regularization view implies that the reconstruction loss itself can be reused as the invariance objective, so TDM adds no auxiliary head or adversarial loss.
  • Using DDIM acceleration, the inference cost drops dramatically relative to DDPM-based DreamRec while keeping performance, which makes generative sequential recommendation practical at scale.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the alignment assumption holds, a general recipe emerges: for any conditional generative model with partially observed conditioning, train by self-inflicted, distribution-preserving deletion of conditioning tokens rather than by imputation.
  • Because the entropy-based stability score identifies sequences where preference is volatile, the same signal could be used at serving time to lower the removal threshold or to demand more evidence before recommending, a use the paper does not explore.
  • The method's success should depend on the quality of the embedding space in which continuity is measured; a testable extension is to pair DTS with content-enriched item embeddings and check whether the gains grow as cosine similarity becomes more semantically meaningful.
  • The paper's own limitation suggests a natural successor: replace the hand-defined continuity and stability metrics with a learned model of preference evolution while keeping the Thompson-sampling framework, and test whether robustness improves further.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. This manuscript proposes TDM, a diffusion-based generative sequential recommender designed to be robust to missing entries in user interaction sequences. Rather than imputing missing interactions, TDM deliberately introduces additional artificial missingness: a dual-side Thompson sampling procedure (Section 4.1) scores items by local continuity and sequences by global stability, and edits sequences by removing high-continuity items from high-stability sequences. The edited sequence is encoded as guidance for a DDIM-based conditional diffusion model trained with classifier-free guidance (Section 4.2). The authors claim that training on such edited sequences provides consistency regularization and that, by extrapolation, the model becomes robust to real missing data. Section 4.3 gives a Taylor-expansion argument intended to bound the effect of real missingness by the effect of simulated missingness. Experiments on YooChoose, KuaiRec, Zhihu, Steam, Beauty, and Toys compare TDM with sequential and generative baselines, including recovering-based methods, and report consistent improvements; ablations and a DTS extension to non-diffusion recommenders are also provided.

Significance. If the central transfer claim were rigorously supported, TDM would be a valuable alternative to recovery-based approaches: it is simple, computationally efficient, and shows consistent improvements across several datasets and baseline families. The paper has concrete strengths: the experimental comparison is broad (Tables 2 and 3), ablations isolate the continuity and stability metrics (Table 4), the DDIM/DDPM comparison (Table 5) is useful, and the code and data are released. The main weakness is that the theoretical justification in Section 4.3 does not actually establish extrapolation from simulated to real missingness; as written, it assumes the key property it sets out to prove. The empirical evidence is encouraging, but the transfer mechanism is not validated under missingness mechanisms that differ from DTS's preference-preserving deletions.

major comments (3)
  1. [Section 4.3, Eq. (19)] The extrapolation argument is circular. The coefficient ratio is bounded only when 'the missing data process from e_{1:N-1} to e_{1:N-1}⊖δ′ can align well with that from e_{1:N-1}⊕δ to e_{1:N-1};' this is exactly the claim that simulated deletions mimic real missing data, and it is not derived from the model or from any stated assumption. The final paragraph of Section 6 acknowledges that real missing data can arise from exposure or popularity bias, which are not preference-preserving processes of the type DTS is designed to emulate. Because the alignment premise is the load-bearing assumption, the section should either be reframed explicitly as intuition or supplied with a formal, testable condition under which the two guidance differences are approximately equal.
  2. [Section 4.3, Eq. (20)] The consistency regularization claim is not supported by the inequality presented. Equation (20) is a trivial consequence of ||a-b||_2^2 ≤ 2(||a-y||_2^2 + ||b-y||_2^2); any two functions that both approximate y satisfy it. The training objective L in Eq. (14) is a sum of per-sample reconstruction losses and contains no term that directly penalizes disagreement between f(e,ĝ) and f(e,g̃). Hence the text does not show that L 'achieves consistency regularization' in the sense of enforcing invariance to simulated missingness. If consistency is a stated contribution, the authors should either add an explicit consistency penalty and ablate it, or weaken the wording to describe a property of the objective.
  3. [Section 5.5] The synthetic missingness experiment does not establish transfer to real missingness as described. The text reports creating datasets with 10-30% missing data but does not specify the deletion mechanism. If the deletion is uniform random, it is a different process from DTS and, more importantly, it does not test the claimed extrapolation for non-preference-preserving mechanisms (e.g., exposure or popularity bias). Please specify the mechanism and include at least one condition in which missingness is correlated with item popularity or recency, or provide an evaluation with naturally missing observations from a dataset with a known complete ground truth.
minor comments (4)
  1. [Tables 2 and 3] The 'Improv.' rows are computed relative to TDM's own score rather than the baseline score. For example, on YooChoose HR the table reports 9.85% for TDM over PDRec, which is (6.90−6.22)/6.90; the standard relative improvement is (6.90−6.22)/6.22 ≈ 10.93%. Please recompute all improvement percentages.
  2. [Abstract, Section 5, and Table 1] The paper says 'five datasets,' but Table 1 lists six datasets (YooChoose, KuaiRec, Zhihu, Steam, Beauty, and Toys); the counts should be made consistent.
  3. [Section 4.1.2 and Section 3.2] The procedure in Eqs. (9)-(11) uses fixed thresholds λ1 and λ2 after drawing probabilities, with no posterior update or exploration-exploitation tradeoff as described for Thompson sampling in Section 3.2. Either define the sampling distribution F(v,p) concretely or rename the procedure to avoid the mismatch.
  4. [Section 5.1] No statistical significance tests are reported for the main comparisons. Given the reported standard deviations, pairwise significance tests (e.g., paired bootstrap) would make the improvement claims more convincing.

Circularity Check

1 steps flagged · score 5.0 of 10

Extrapolation proof in Sec. 4.3 assumes the alignment it needs to prove; the empirical evaluation is not circular.

  1. other [Section 4.3, Extrapolation, bound of coefficient after Eq. (19)]
    "If we can simulate the mechanism of missing data — specifically, if the missing data process from e1:N−1 to e1:N−1⊖δ′ can align well with that from e1:N−1⊕δ to e1:N−1 — the difference between the two pair of data will be roughly equivalent. Consequently, the two differences in guidance, ḡ−ĝ and ĝ−g̃, will also be approximately equal. In this scenario, the coefficient will be close to 1, resulting in a bounded value C > 0. Thus we can validate that enhancing diffusion models' insensitivity to simulated missing data enables resilience against real missing data."

    The extrapolation theorem is the paper's central justification that training on DTS-simulated deletions transfers to real missing data. The only step that bounds the coefficient in Eq. (19) is the assumption that the simulated missingness process from observed to edited sequences aligns with the real missingness process from complete to observed sequences. That alignment is precisely the substantive property TDM is claimed to provide; the preceding description of DTS says it preserves user preference evolution, but that is not shown to imply alignment with real missingness mechanisms such as exposure bias or popularity bias, which the paper's own Conclusion acknowledges.

full rationale

The paper's empirical contribution is not circular: TDM is trained on sequences edited by DTS and evaluated on standard benchmarks and synthetic missing-data versions, with improvements over strong baselines reported. The consistency-regularization inequality (Eq. 20) is a valid though loose bound, and calling it a regularization effect is not circular because the reconstruction terms are genuinely minimized during training. The main circularity concern is the theoretical extrapolation proof in Section 4.3. Equation (19) is derived correctly by Taylor expansion, but the crucial coefficient bound is obtained only by assuming that the simulated missingness process aligns with the real missingness process. This alignment is the very property that would make the method work, so the proof does not provide independent support for the extrapolation claim; it reduces the 'validation' to an unverified premise. Because the experimental evaluation is self-contained and the theoretical issue is localized, the overall circularity is partial rather than total.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The method depends on two tuned thresholds and a guidance weight, all chosen on validation. The theoretical justification rests on several unproven modeling assumptions, most importantly that DTS-simulated gaps resemble real gaps. No new physical entities or mediators are introduced.

free parameters (5)
  • lambda_1 = not reported; tuned in [0, 0.1, ..., 1]
    Threshold deciding whether a sequence is edited in DTS (Eq. 10). Chosen on validation.
  • lambda_2 = not reported; tuned in [0, 0.1, ..., 1]
    Threshold deciding whether an item is removed (Eq. 11). Chosen on validation.
  • w = not reported; tuned in [0, 2, 4, 6, 8, 10]
    Guidance strength in classifier-free guidance (Eq. 15). Chosen on validation.
  • T = not reported; tuned in [500, 1000, 2000]
    Total diffusion steps for the variance schedule.
  • tau_S = not reported; tuned in [5, 10, 20]
    Number of DDIM reverse steps used at inference.
assumptions (5)
  • standard math The denoising network f_theta is differentiable and Taylor's formula applies to it.
    Used in Eq. 17 to expand f_theta around the observed guidance g_hat.
  • domain assumption High-continuity items in high-stability sequences carry little preference information, so removing them preserves the preference evolution.
    Core to DTS; stated in Section 4.1.2 and illustrated in Figure 1, but not formally validated.
  • ad hoc to paper The simulated missingness via DTS matches the real missingness process, so that (bar_g minus hat_g) and (hat_g minus tilde_g) are approximately equal.
    Needed in Section 4.3 to bound the extrapolation coefficient; this is exactly the property the paper aims to achieve, making the proof circular.
  • domain assumption User preference evolution is adequately captured by continuity scores (adjacent-item similarity) and sequence-level entropy stability.
    Underlies both probability models in Section 4.1.1.
  • domain assumption The temporal 8:1:1 split prevents future interactions from leaking into training.
    Experimental protocol following prior work, used to ensure valid offline evaluation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Addressing Missing Data Issue for Diffusion-based Recommendation." pith.science (2026). https://pith.science/paper/S53RUHGX

@misc{pith2026250512283,
  author       = {Pith},
  title        = {Pith review of: Addressing Missing Data Issue for Diffusion-based Recommendation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S53RUHGX}},
  note         = {Machine review of arXiv:2505.12283}
}
read the original abstract

Diffusion models have shown significant potential in generating oracle items that best match user preference with guidance from user historical interaction sequences. However, the quality of guidance is often compromised by unpredictable missing data in observed sequence, leading to suboptimal item generation. Since missing data is uncertain in both occurrence and content, recovering it is impractical and may introduce additional errors. To tackle this challenge, we propose a novel dual-side Thompson sampling-based Diffusion Model (TDM), which simulates extra missing data in the guidance signals and allows diffusion models to handle existing missing data through extrapolation. To preserve user preference evolution in sequences despite extra missing data, we introduce Dual-side Thompson Sampling to implement simulation with two probability models, sampling by exploiting user preference from both item continuity and sequence stability. TDM strategically removes items from sequences based on dual-side Thompson sampling and treats these edited sequences as guidance for diffusion models, enhancing models' robustness to missing data through consistency regularization. Additionally, to enhance the generation efficiency, TDM is implemented under the denoising diffusion implicit models to accelerate the reverse process. Extensive experiments and theoretical analysis validate the effectiveness of TDM in addressing missing data in sequential recommendations.

Figures

Figures reproduced from arXiv: 2505.12283 by the authors.

Figure 1
Figure 1. Phenomenon of uncertain missing data in se [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The overview of the TDM framework, which simulates extra missing data with DTS in the guidance signals, achieving [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Sensitivity of TDM to the hyperparameter of [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Sensitivity of TDM to the hyperparameter of [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Performance of TDM on synthetic datasets with different missing ratios. [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Performance of TDM on multiple datasets with different sequence lengths. [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

65 extracted references · 46 canonical work pages

  1. [1]

    David Ben-Shimon, Alexander Tsikinovsky, Michael Friedmann, Bracha Shapira, Lior Rokach, and Johannes Hoerle. 2015. RecSys Challenge 2015 and the YOO- CHOOSE Dataset. In RecSys. ACM, 357–358

    work page 2015

  2. [2]

    Ziqiang Cui, Haolun Wu, Bowei He, Ji Cheng, and Chen Ma. 2024. Diffusion-based Contrastive Learning for Sequential Recommendation. CoRR abs/2405.09369 (2024)

    arXiv 2024

  3. [3]

    Xinyan Fan, Jianxun Lian, Wayne Xin Zhao, Zheng Liu, Chaozhuo Li, and Xing Xie. 2022. Ada-Ranker: A Data Distribution Adaptive Ranking Paradigm for Sequential Recommendation. In SIGIR. ACM, 1599–1610

    work page 2022

  4. [4]

    Ziwei Fan, Zhiwei Liu, Yu Wang, Alice Wang, Zahra Nazari, Lei Zheng, Hao Peng, and Philip S. Yu. 2022. Sequential Recommendation via Stochastic Self-Attention. In WWW. ACM, 2036–2047

    work page 2022

  5. [5]

    Raymond Feng, Flávio Calmon, and Hao Wang. 2023. Adapting Fairness Inter- ventions to Missing Values. In NeurIPS

    work page 2023

  6. [6]

    Chongming Gao, Shijun Li, Wenqiang Lei, Jiawei Chen, Biao Li, Peng Jiang, Xiangnan He, Jiaxin Mao, and Tat-Seng Chua. 2022. KuaiRec: A Fully-observed Dataset and Insights for Evaluating Recommender Systems. In CIKM. ACM, 540–550

    work page 2022

  7. [7]

    Bin Hao, Min Zhang, Weizhi Ma, Shaoyun Shi, Xinxing Yu, Houzhi Shan, Yiqun Liu, and Shaoping Ma. 2021. A Large-Scale Rich Context Query and Recommen- dation Dataset in Online Knowledge-Sharing. CoRR abs/2106.06467 (2021)

    work page Pith review arXiv 2021

  8. [8]

    Balázs Hidasi, Alexandros Karatzoglou, Linas Baltrunas, and Domonkos Tikk

Show all 65 references
  1. [9]

    Erna Hikmawati, Heru Nugroho, and Kridanto Surendro. 2024. Improve the Quality of Recommender Systems based on Collaborative Filtering with Missing Data Imputation. In ICSCA. ACM, 75–80

    2024

  2. [10]

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. 2020. Denoising Diffusion Probabilistic Models. In NeurIPS

    2020

  3. [11]

    Jonathan Ho and Tim Salimans. 2022. Classifier-Free Diffusion Guidance. In NeurIPS Workshop on Deep Generative Models and Downstream Applications

    2022

  4. [12]

    Yitong Ji, Aixin Sun, Jie Zhang, and Chenliang Li. 2023. A Critical Study on Data Leakage in Recommender System Offline Evaluation. ACM Trans. Inf. Syst. 41, 3 (2023), 75:1–75:27

    2023

  5. [13]

    Yangqin Jiang, Yuhao Yang, Lianghao Xia, and Chao Huang. 2024. DiffKG: Knowledge Graph Diffusion Model for Recommendation. In WSDM. ACM, 313– 321

    2024

  6. [14]

    Wang-Cheng Kang and Julian J. McAuley. 2018. Self-Attentive Sequential Rec- ommendation. In ICDM. IEEE Computer Society, 197–206

    2018

  7. [15]

    Xiaoyu Kong, Jiancan Wu, An Zhang, Leheng Sheng, Hui Lin, Xiang Wang, and Xiangnan He. 2024. Customizing Language Models with Instance-wise LoRA for Sequential Recommendation. In NeurIPS

    2024

  8. [16]

    Wuchao Li, Rui Huang, Haijun Zhao, Chi Liu, Kai Zheng, Qi Liu, Na Mou, Guorui Zhou, Defu Lian, Yang Song, Wentian Bao, Enyun Yu, and Wenwu Ou. 2025. DimeRec: A Unified Framework for Enhanced Sequential Recommendation via Generative Diffusion Models. In WSDM. ACM, 726–734

    2025

  9. [17]

    Zihao Li, Aixin Sun, and Chenliang Li. 2024. DiffuRec: A Diffusion Model for Sequential Recommendation. ACM Trans. Inf. Syst. 42, 3 (2024), 66:1–66:28

    2024

  10. [18]

    Zongwei Li, Lianghao Xia, and Chao Huang. 2024. RecDiff: Diffusion Model for Social Recommendation. In CIKM. ACM, 1346–1355

    2024

  11. [19]

    Zhi Li, Hongke Zhao, Qi Liu, Zhenya Huang, Tao Mei, and Enhong Chen. 2018. Learning from History and Present: Next-item Recommendation via Discrimina- tively Exploiting User Behaviors. In KDD. ACM, 1734–1743

    2018

  12. [20]

    Jiayi Liao, Sihang Li, Zhengyi Yang, Jiancan Wu, Yancheng Yuan, Xiang Wang, and Xiangnan He. 2024. LLaRA: Large Language-Recommendation Assistant. In SIGIR. ACM, 1785–1795

    2024

  13. [21]

    Yujie Lin, Chenyang Wang, Zhumin Chen, Zhaochun Ren, Xin Xin, Qiang Yan, Maarten de Rijke, Xiuzhen Cheng, and Pengjie Ren. 2023. A Self-Correcting Sequential Recommender. In WWW. ACM, 1283–1293

    2023

  14. [22]

    Qidong Liu, Fan Yan, Xiangyu Zhao, Zhaocheng Du, Huifeng Guo, Ruiming Tang, and Feng Tian. 2023. Diffusion Augmentation for Sequential Recommendation. In CIKM. ACM, 1576–1586

    2023

  15. [23]

    Haokai Ma, Ruobing Xie, Lei Meng, Xin Chen, Xu Zhang, Leyu Lin, and Zhanhui Kang. 2024. Plug-In Diffusion Model for Sequential Recommendation. In AAAI. AAAI Press, 8886–8894

    2024

  16. [24]

    Wenyu Mao, Shuchang Liu, Haoyang Liu, Haozhe Liu, Xiang Li, and Lanatao Hu. 2025. Distinguished Quantized Guidance for Diffusion-based Sequence Recommendation. CoRR abs/2501.17670 (2025)

    2025 arXiv

  17. [25]

    Wenyu Mao, Jiancan Wu, Weijian Chen, Chongming Gao, Xiang Wang, and Xiangnan He. 2025. Reinforced prompt personalization for recommendation with large language models. ACM Trans. Inf. Syst. 43, 3 (2025), 1–27

    2025

  18. [26]

    McDonald, Lucas Maystre, Mounia Lalmas, Daniel Russo, and Kamil Ciosek

    Thomas M. McDonald, Lucas Maystre, Mounia Lalmas, Daniel Russo, and Kamil Ciosek. 2023. Impatient Bandits: Optimizing Recommendations for the Long- Term Without Delay. In KDD. ACM, 1687–1697

    2023

  19. [27]

    Hall, Daniel Cer, and Yinfei Yang

    Jianmo Ni, Gustavo Hernández Ábrego, Noah Constant, Ji Ma, Keith B. Hall, Daniel Cer, and Yinfei Yang. 2022. Sentence-T5: Scalable Sentence Encoders from Pre-trained Text-to-Text Models. In ACL (Findings). Association for Computa- tional Linguistics, 1864–1874

    2022

  20. [28]

    Yong Niu, Xing Xing, Zhichun Jia, Ruidi Liu, Mindong Xin, and Jianfu Cui

  21. [29]

    Ian Osband and Benjamin Van Roy. 2015. Bootstrapped Thompson Sampling and Deep Exploration. CoRR abs/1507.00300 (2015)

    2015 arXiv

  22. [30]

    William Peebles and Saining Xie. 2023. Scalable Diffusion Models with Trans- formers. In ICCV. IEEE, 4172–4182

    2023

  23. [31]

    Ruihong Qiu, Zi Huang, Hongzhi Yin, and Zijian Wang. 2022. Contrastive Learn- ing for Representation Degeneration Problem in Sequential Recommendation. In WSDM. ACM, 813–823

    2022

  24. [32]

    Tran, Jonah Samost, Maciej Kula, Ed H

    Shashank Rajput, Nikhil Mehta, Anima Singh, Raghunandan Hulikal Keshavan, Trung Vu, Lukasz Heldt, Lichan Hong, Yi Tay, Vinh Q. Tran, Jonah Samost, Maciej Kula, Ed H. Chi, and Mahesh Sathiamoorthy. 2023. Recommender Systems with Generative Retrieval. In NeurIPS

    2023

  25. [33]

    Daniel Russo, Benjamin Van Roy, Abbas Kazerouni, Ian Osband, and Zheng Wen

  26. [34]

    Noveen Sachdeva, Giuseppe Manco, Ettore Ritacco, and Vikram Pudi. 2019. Se- quential Variational Autoencoders for Collaborative Filtering. In WSDM. ACM, 600–608

    2019

  27. [35]

    Jiaming Song, Chenlin Meng, and Stefano Ermon. 2021. Denoising Diffusion Implicit Models. In ICLR. OpenReview.net

    2021

  28. [36]

    Harald Steck. 2013. Evaluation of recommendations: rating-prediction and rank- ing. In RecSys. ACM, 213–220

    2013

  29. [37]

    Fei Sun, Jun Liu, Jian Wu, Changhua Pei, Xiao Lin, Wenwu Ou, and Peng Jiang

  30. [38]

    Jiaxi Tang and Ke Wang. 2018. Personalized Top-N Sequential Recommendation via Convolutional Sequence Embedding. In WSDM. ACM, 565–573

    2018

  31. [39]

    Chenxu Wang, Fuli Feng, Yang Zhang, Qifan Wang, Xunhan Hu, and Xiangnan He

  32. [40]

    Jun Wang, Haoxuan Li, Chi Zhang, Dongxu Liang, Enyun Yu, Wenwu Ou, and Wenjia Wang. 2023. CounterCLR: Counterfactual Contrastive Learning with Non-random Missing Data in Recommendation. In ICDM. IEEE, 1355–1360

    2023

  33. [41]

    Wenjie Wang, Yiyan Xu, Fuli Feng, Xinyu Lin, Xiangnan He, and Tat-Seng Chua

  34. [42]

    Xiaojie Wang, Rui Zhang, Yu Sun, and Jianzhong Qi. 2019. Doubly Robust Joint Learning for Recommendation on Data Missing Not at Random. In ICML (Proceedings of Machine Learning Research, Vol. 97) . PMLR, 6638–6647

    2019

  35. [43]

    Zhendong Wang, Yifan Jiang, Huangjie Zheng, Peihao Wang, Pengcheng He, Zhangyang Wang, Weizhu Chen, and Mingyuan Zhou. 2023. Patch Diffusion: Faster and More Data-Efficient Training of Diffusion Models. In NeurIPS

    2023

  36. [44]

    Zhenlei Wang, Shiqi Shen, Zhipeng Wang, Bo Chen, Xu Chen, and Ji-Rong Wen

  37. [45]

    Zhendong Wang and Mingyuan Zhou. 2020. Thompson Sampling via Local Uncertainty. In ICML (Proceedings of Machine Learning Research, Vol. 119) . PMLR, 10115–10125

    2020

  38. [46]

    Xu Xie, Fei Sun, Zhaoyang Liu, Shiwen Wu, Jinyang Gao, Jiandong Zhang, Bolin Ding, and Bin Cui. 2022. Contrastive Learning for Sequential Recommendation. In ICDE. IEEE, 1259–1273

    2022

  39. [47]

    In SIGIR

    Diffusion Recommender Model. In SIGIR. ACM, 832–841

  40. [48]

    Yiyan Xu, Wenjie Wang, Fuli Feng, Yunshan Ma, Jizhi Zhang, and Xiangnan He

  41. [49]

    Zhengyi Yang, Jiancan Wu, Zhicai Wang, Xiang Wang, Yancheng Yuan, and Xiangnan He. 2023. Generate What You Prefer: Reshaping Sequential Recom- mendation via Guided Diffusion. In NeurIPS

    2023

  42. [50]

    Chi Zhang, Qilong Han, Rui Chen, Xiangyu Zhao, Peng Tang, and Hongtao Song

  43. [51]

    Han Zhang, Zizhao Zhang, Augustus Odena, and Honglak Lee. 2020. Consistency Regularization for Generative Adversarial Networks. In ICLR. OpenReview.net

    2020

  44. [52]

    Jujia Zhao, Wenjie Wang, Yiyan Xu, Teng Sun, Fuli Feng, and Tat-Seng Chua

  45. [53]

    Xiaolin Zheng, Menghan Wang, Renjun Xu, Jianmeng Li, and Yan Wang. 2022. Modeling Dynamic Missingness of Implicit Feedback for Sequential Recommen- dation. IEEE Trans. Knowl. Data Eng. 34, 1 (2022), 405–418

    2022

  46. [54]

    Da Xu, Chuanwei Ruan, Evren Körpeoglu, Sushant Kumar, and Kannan Achan

  47. [57]

    In SIGIR

    Diffusion Models for Generative Outfit Recommendation. In SIGIR. ACM, 1350–1359

  48. [60]

    SSDRec: Self-Augmented Sequence Denoising for Sequential Recommen- dation. In ICDE. IEEE, 803–815

  49. [63]

    In SIGIR

    Denoising Diffusion Recommender Model. In SIGIR. ACM, 1370–1379

  50. [65]

    Yunqin Zhu, Chao Wang, Qi Zhang, and Hui Xiong. 2024. Graph Signal Diffusion Model for Collaborative Filtering. In SIGIR. ACM, 1380–1390

    2024

  51. [2016]

    In ICLR (Poster)

    Session-based Recommendations with Recurrent Neural Networks. In ICLR (Poster)

  52. [2018]

    A Tutorial on Thompson Sampling. Found. Trends Mach. Learn. 11, 1 (2018), 1–96

    2018

  53. [2019]

    BERT4Rec: Sequential Recommendation with Bidirectional Encoder Repre- sentations from Transformer. In CIKM. ACM, 1441–1450

  54. [2020]

    In NeurIPS

    Adversarial Counterfactual Learning and Evaluation for Recommender System. In NeurIPS

  55. [2022]

    Unbiased Sequential Recommendation with Latent Confounders. In WWW. ACM, 2195–2204

  56. [2023]

    IEEE Trans

    Rethinking Missing Data: Aleatoric Uncertainty-Aware Recommendation. IEEE Trans. Big Data 9, 6 (2023), 1607–1619

    2023

  57. [2024]

    In WWW (Companion Volume)

    Diffusion Recommendation with Implicit Sequence Influence. In WWW (Companion Volume). ACM, 1719–1725

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.